基于人工智能斑块分割超声图像的影像组学在颈动脉斑块稳定性评估中的应用

目的 探讨基于人工智能斑块分割超声图像的影像组学评估颈动脉斑块的稳定性,以及其对颈动脉易损斑块与稳定性斑块的诊断效能.方法 171例缺血性脑卒中患者通过颈动脉彩超检查分为易损斑块组(83例)与稳定斑块组(88例).在人工智能斑块分割超声图像上提取369个影像组学特征,采用最小绝对收缩和选择算子方法 对影像组学特征进行降...

Application of radiomics based on artificial intelligence in evaluation of carotid plaque stability

Objective To evaluate the stability of carotid atherosclerotic plaques and its diagnostic efficacy for vulnerable and stable carotid atherosclerotic plaques by using radiomics based on ultrasound image segmentation with artificial intelligence.Methods 171 patients with ischemic stroke were divided into vulnerable plaque group(83 cases)and stable plaque group(88 cases)by carotid ultrasound examination.369 imageomic features were extracted from ultrasound image of artificial intelligence plaque segmentation.The minimum absolute shrinkage and selection operator method were used to reduce and screen the imageomic features.The selected variables were further modeled and verified by support vector machine(SVM).The sensitivity,specificity and area under the curve(AUC)of the model were evaluated by receiver operating characteristic(ROC)curve analysis.Results A total of 21 imaging characteristic parameters were selected.The AUC of training group was 0.984(95%confidence interval:0.971-0.997),sensitivity was 92.0%,specificity was 95.2%.The AUC of validation group was 0.964(95%confidence interval:0.935-0.993),sensitivity was 100.0%,specificity was 84.8%.Conclusion The image omics model based on artificial intelligence plaque segmentation ultrasound image can effectively distinguish vulnerable plaque and stable plaque,which provides a new method for rapid and accurate evaluation of the stability of carotid plaque.